How to get started with deep learning using MRI data.
Even though all the algorithms and information are open-source these days, sometimes using even the most well-established computer vision or deep learning methods do not produce expected results, especially for medical imaging problems. The problem lies in the insufficient understanding of medical data and its efficient use to leverage the power of the new computational methods. This post here addresses a basic structure that can help in understanding the problem at hand and implement deep learning models to use MRI data. Although this work primarily deals with classification problems, the data exploration and preparation steps equally apply to other types of problem statements. Before we start, it is beneficial to know that most of the popular machine learning libraries and deep learning frameworks used here based on python. The requirements list below contains the language and library requirements necessary for following the code of this tutorial.
Requirements: Main libraries and python version used for the code mentioned in this tutorial are as follows:
- Python: 3.6
- SimpleITK : 1.2.4
- torch (PyTorch library): 1.4.0
- torchvision (datasets and transforms): 0.5.0
- sklearn: 0.0
- Understanding Magnetic Resonance Imaging (MRI) and DICOM
Magnetic Resonance Imaging (MRI) is a widely used imaging modality in radiology. Even though MRI has long data…
